Hal Daume, Volpi-Cupal Family Endowed Professor in the Department of Computer Science and Director of the Artificial Intelligence Interdisciplinary Institute at Maryland (AIM)
The future of Spatial Intelligence - The opportunity for Leadership In the not-too-distant future, every decision - investment, infrastructure, climate adaptation, real estate, farming, banking - will start with a live conversation with the Earth. In our field, we have surely moved from looking at the Earth to asking it questions building on the maturity of geospatial technology and computing, the explosion of earth observation data, and the proliferation of AI. In an era of 'AI Everywhere,' the path to true spatial intelligence requires more than just technological advancement; it demands a radical shift in how we collaborate.
In her keynote, Marge Cole draws on her global experience working with NASA, OGC and a multitude of startups and innovators over the years. Reflecting on the path towards spatial intelligence, its impact on innovation, research, and business opportunties - underscoring the pivotal need for academia and research to forge more cross-disciplinary, more collaboration, more agility, and more partnerships with industry upfront and throughout the research process, also examining the unique opportunities and challenges this rapid evolution presents for education and research.
Driven by a passion for innovation that empowers, I have spent over two decades at the intersection of aerospace, geospatial standards, and entrepreneurship. From my tenure at NASA, SGT and KBR to launching my own consulting firm and now LunateAI, I have a proven track record of championing... Read More →
Cluster 1: Urban Perception & Human-Centered AI Keenon Lindsey, Texas State Univ.: “Seeing” Gentrification: A Deep Learning Approach to Visual Change Perception Yingrui Zhao, Univ. of Maryland: An LLM-Guided Approach for Analyzing Public Sentiment associated with Transportation POI Visit Patterns Michaelmary Chukwu, Univ. of Maryland: From Gravity Models to Semantic Reasoning: Leveraging LLMs for Visual Destination Characterization
Cluster 2: Environment, Hazards & Remote Sensing Sandra Le, George Mason Univ.: A Spatiotemporal Analysis of Vegetation and Water Changes in Libya Extreme Rainfall 2023 Using Remote Sensing Products and the Google Earth Engine (GEE) Xin Dong, Univ. of Maryland: Predicting the spatio-temporal spread of Plasmodium vivax malaria using a human-movement–informed GeoAI model Aleksander Berg, Univ. of Colorado Boulder: Using Foundation Model Embeddings to Map Colorado's Built Hazard Interface for Wildfire
— Break / Reset (5 min.) —
Cluster 3: GeoAI Methods & Modeling Jikun Liu, Texas A&M Univ.: A Wide-and-Deep-Based Time Sequence Model for Predicting Power Outages Caused by Extreme Winter Storms (canceled) Victor Irekponor, Univ. of Maryland: Text-to-Visualization for Spatially Varying Coefficient Models: Encoding SVC Visualization Principles in Language-Driven Workflows Zhihao Wang, Univ. of Maryland: TreeFinder: AI Everywhere in Forest Monitoring — A National-Scale GeoAI Benchmark for Individual Tree Mortality
Cluster 4: Spatial Theory & Advanced Methods Mengyu Liao, Univ. of Maryland: Change of Support as a Reasoning Layer in LLM-Based GIS Workflows Jina Kim, Univ. of Minnesota: Spatial Heterogeneity-Aware Cross-Indicator Transfer for Prediction in Label-Sparse Regions Yuán Niú, Texas A&M University: Disentangling and Tackling the Spatiotemporal Biases in Social Sensing Data: A Cognitive-behavioral Approach
Pick up your box lunch and come to the Atrium Lounge!
As the 2026 midterm elections approach, gerrymandering has once again entered public and scholarly debate. For GIScience (GISc), political geography, and spatial justice research, gerrymandering has never been simply a technical question of how to draw district lines. It is about how spatial boundaries intentionally shape political representation, community (non)visibility, public resource allocation, and the ways people are governed and made visible. GISc has transformed redistricting, where GISc is a “double-edged sword” for gerrymandering. GISc can be used to create highly targeted gerrymanders, but it can also be used to detect and challenge them. Moreover, in the age of AI Everywhere, this long-standing issue is becoming even more complex: AI may not only change how redistricting and boundary manipulation are carried out but also reshape the spatial logic of political governance itself.
This discussion will begin with the impact of GISc and AI on gerrymandering and redistricting. In recent years, machine learning, nonparametric statistical learning, and algorithm-assisted redistricting have been used to generate and evaluate alternative districting plans, helping identify anomalous bias, explain district structures, and improve transparency in redistricting analysis (Stolicki et al., 2024). At the same time, we need to ask: Could AI also be used to create more refined and less visible forms of boundary manipulation? If district boundaries have already distorted community representation and public data, might AI-mediated governance further amplify these distortions? In the age of AI Everywhere, is gerrymandering expanding from a manipulation of electoral boundaries into a spatial infrastructure problem that shapes data, representation, public resources, and algorithmic governance? And how might we use GeoAI not only to detect manipulation, but also to advance spatial justice, democratic representation, and responsible AI governance?
Samrin Sauda, Penn State: Improving Compound Coastal Flood Forecasting with Deep Learning Along the Southeast U.S. Coast
Kendall Phillips, University Of Maine: Comparison of PFAS Contamination Across Drinking Water Wells in Maine Using a Knowledge Graph Approach
Madhukar Kuchavaram, Univ. of Florida: Spatiotemporal Forecasting for Proactive Vector Control: Weather-Lagged Prediction Models for Mosquito Surveillance in Corpus Christi, Texas
Yuán Niú, Texas A&M Univ.: A Spatial Decision-Support Tool to Enhance Participatory Planning for Urban Heat Resilience
Muhammad Khattak, Auburn University: SMARTS: A Synchronous Multi-Agent Reinforcement Learning-driven Transit Scheduler via Graph Neural Networks and Proximal Policy Optimization
Emily Zhou, University of California Santa Barbara: Evaluating Methodological Structure and Analytical Outcome Divergence Across Machine Learning and Multi-Criteria Decision Analysis Approaches in Spatial Decision Support Research
Paul C. Dunn, Oregon State Univ.: Retrieval-Augmented 4D Visualization for a Digital Twin of the Ocean: Improving Multiscale Pattern–Process Analysis with Generative AI, Reinforcement Learning, and Heterogeneous I/O
Omada Friday Ojonugwa, Beihang Univ.: Multi-Sensor Fusion for Soil Moisture Estimation in West Africa using Ensemble Learning
Shirin Alsadat Mahmoudian, George Mason Univ.: Generative AI for Urban Infrastructure Auditing: A Micro-scale Evaluation of Multimodal Transit Environments
Isaac Quaye, Temple Univ.: Explainability of DL-based Gentrification Detection from Street View Imagery
Jiahua Chen, University of California, Santa Barbara: Making POI Uncertainty Legible: A Pipeline-Aware Error Framework and Audit Design for Urban Analytics
Zhongqi Zheng, University of California, Santa Barbara: Pixels to Paths: Spatially Informed Computer Vision for Road Network Topology Assessment
Hailey Richardson, Univ. of Alabama: Understanding the Impact of Spatial Relationship Definitions on Crime Clustering Analysis: Implications for Urban Crime Patterns and Policing Strategies
Youshuang Hu, Univ. of Connecticut: GIScience for Social Equity: Unpacking the Black Box of Spatial Inequality
Jon Nealon, SUNY Albany: Beyond the God's Eye View: The Hot Air Balloon Perspective in Geospatial Journalism
We are seeing changes in the NSF and NIH (etc.) funding landscape. This has and may continue to result in job cuts, terminations of grants, loss of overhead, and dissolution of programs. GIScientists may benefit from an open discussion about the current research climate.
This is a very informal sharing and listening session on reactions to and strategies for managing the changes in federal funding allocation (both toward and away different priorities) and is open to all. Participants can get new ideas for research support, new knowledge about how institutions are coping, and clarity about proposed legislation and effects.
There is no one speaker or representative from any agency leading or joining the discussion, and we will depend on those who come to the session to share their experiences.
Potential topics of discussion: -How NSF-HEGS (formerly GSS) and related programs that sponsor GIScience research have helped us. -Experiences with NSF funding for GIScientists through a variety of programs. -How academic departments and administrations are managing funding changes. -Ideas for a future webinar or other future directions to continue the conversation. -Other topics of interest.
Come join us for our UCGIS Symposium poster session and reception at the Samuel Riggs IV Alumni Center. We have a great lineup of posters! Get caught up with other UCGIS members at this not-to-missed social event. This Center is just a short walk (10-12 mins, mostly a gentle downhill) from the ESJ building where Symposium sessions are being held. We will have shuttles to take you back to The Hotel, Cambria, and Marriott towards the end of the reception.
Students Jayanta Biswas, UNC Charlotte: A Deep Learning Framework for Fusing Multi-Modal Environmental Data to Downscale Human Mobility for Precision Malaria Modeling in Zambia Arati Budhathoki, Clemson Univ.: Tracking Mountain Degradation for the United Nations (UN) Sustainable Development Goals (SDGs) Using the State of Colorado (USA) as an Example Sofiia Drozd, NTUU KPI: Agentic AI Framework for Automated Mapping of War-Damaged Agricultural Risk Zones and Satellite Data Retrieval Paul C. Dunn, Oregon State Univ.: Improving Multiscale Pattern–Process Analysis with a Eulerian-Lagrangian Flow Model and Uncertainty Aware Clustering Analysis Maxwell Gundling, Salisbury Univ.: From Silos to Spatial Data: An Enterprise GIS for Historical Research Fatemeh Janatabadi, George Mason Univ.: Artificial Intelligence Drives a New Feedback Loop Between Human Mobility and Urban Landscapes SiyuLu,Texas A&M Univ.: Deep Learning versus Traditional Interpolation for Elevation Reconstruction: Evaluating Performance Gains from Terrain-Based Auxiliary Variables Oliver Matus-Bond, Macalester College: Mapping the spatial relationship between invasive Melaleuca quinquenervia and fire occurrence in southeastern Madagascar Haley Mullen, Univ. of Maryland: LLM-based generation of geospatial synthetic data for predicting chronic disease Hossein Naderi, Texas A&M Univ.–Corpus Christi: Using Large Language Models to Quantify Urban Environments from Google Street View Mahsa Saharkhiz, Univ. at Buffalo: The Impact of Form-Based Zoning on Residential Values: A Geospatial Analysis of Buffalo's Green Code (2013–2024) Zahra Salehi, Univ. of Connecticut: Spatial Intelligence for Agrivoltaic Land Suitability: A GIS-Based Multi-Criteria Decision Framework in Connecticut Rachel Simon, Salisbury Univ.: From Surface to Subsurface: Mapping Cemeteries in Dorchester County, Maryland Daryna Skakun, Urbana High School: Agentic AI for Environmental Impact Assessment of Construction Projects Using Satellite Data Ruichen Wang, Univ. of Maryland: Coincident Data Discovery Engine (CoDD): Enabling Global Cross-Platform Satellite Data Discovery Zhihao Wang, Univ. of Maryland: CarbonGlobe: A Global ML-Ready Benchmark for Long-Term Carbon Forecasting Under Climate Change
Faculty & Other Wataru Morioka, Salisbury Univ.: Spatial Thinking–Centered GIS Curriculum: Problem Solving, Collaboration, and AI Era Pedagogy at Salisbury University
